Muse Image
Meta's agentic image model, which plans and checks its work before rendering and can ground images in search results. It is priced for volume at $0.01 per image on the Meta Model API.
MetaImage modelReleased July 2026Data checked October 18, 2026
Where people rank it
#4
Editing images (tied), 142,644 votes
#5
Creating images (tied), 42,262 votes
From blind votes on LMArena, where people pick the better of two anonymous answers. See the full tables on the AI leaderboard.
Strengths and weaknesses
- Cheapest image model in this list at $0.01 per image
- Top ten on LMArena text-to-image (9th of 82 entries)
- Available through the Meta Model API, fal, Runway and OpenRouter
- Clearly behind the top four image models on LMArena
- Meta's model page gives no release date or detailed specs
Best for
- Very cheap image generation at scale
- Images grounded in current information
Every score we have
Each bar compares this model with the best result any model on this page has on the same test.
- LMArena text to image1,274Arena score 1274.0 (95% CI 1268.9 to 1279.1); 42,262 votes; entry "muse-image"; position 9 of 82 on the full board
- LMArena image editing1,402.6Arena score 1402.6 (95% CI 1398.4 to 1406.9); 142,644 votes; entry "muse-image"; position 8 of 59 on the full board
- Artificial Analysis image arena1,127Elo 1127.0 (95% CI 1115.0 to 1139.0); 8,253 appearances; position 8 of 104 on the Artificial Analysis text-to-image arena
- Price per 1,000 images (estimate)$10$10.00 per 1,000 images (Artificial Analysis estimate)
Compare with
- Grok Imagine Image 2.0SpaceXSI's (formerly xAI) second-generation Grok Imagine image model for generating and editing images from text or image prompts.
- MAI-Image-2.6Microsoft AI's image model, launched on August 10, 2026 and offered in public preview on Microsoft Foundry.
- Seedream 5.0 ProByteDance Seed's image creation and editing model, launched on July 8, 2026, with region editing, layer splitting and dense infographic text.
- GPT-Image-2.5 SunburstOpenAI's most capable image generation and editing model, released as a September 8, 2026 snapshot alongside the faster GPT-Image-2.5 Flare.